greenstone.com.au — AI Site Grade

Analysis

I have enough data. Let me compile the audit.

---

Greenstone Financial Services: The AI Identity Crisis

The cold LLM knowledge of "Greenstone Australia" describes a gold mining company — a completely different entity — while the actual site is one of the largest direct insurance distributors in the South Pacific, yet it has near-zero external search footprint and no structured data beyond a single homepage Organization schema.

Crawler Access

All major AI crawlers — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, anthropic-ai, Bytespider, Applebot-Extended, ChatGPT-User — receive a 200 with full HTML content identical to browser delivery. The site runs on a CDN (Incapsula/Imperva, via x-iinfo and x-cdn headers) with HSTS and X-Frame-Options: SAMEORIGIN, but no bot-specific blocking. The robots.txt has a catch-all * rule disallowing only /lp/, /search, /ping, and /site-maintenance — no AI-bot-specific directives exist. /llms.txt returns a 404 (an HTML error page, not a plaintext file), meaning no AI-friendly content map is published.

Cold-Knowledge Gap

When queried cold, the LLM describes Greenstone Australia as a gold exploration company focused on the Youanmi Gold Mine in Western Australia — a junior miner with project financing delays. The actual Greenstone is a financial services holding company founded in 2007, operating 13+ insurance brands (Real Insurance, Australian Seniors, Choosi, Guardian Insurance, Buddy Pet Insurance, etc.) across Australia, New Zealand, and Canada, majority-owned by Ontario Teachers' Pension Plan and CDPQ. The model's prior has zero overlap with the site's actual content. This is a severe identity collision: any AI answering "What is Greenstone Australia?" without live retrieval will describe a gold miner, not an insurer.

Schema Posture

Only the homepage carries JSON-LD — a single Organization schema with name, contact point, awards, and a list of 13 sub-organizations. Every other page on the site (about, brands-partners, news, contact, privacy, careers, Canada) has zero structured data. No WebPage, BreadcrumbList, FAQPage, Article, Product, or InsuranceProduct schema exists anywhere. The brands-partners pages describe specific insurance products (life, funeral, pet, car, home) with detailed feature lists but no schema markup to help AI engines understand them as insurance offerings.

External Signals

DuckDuckGo searches for "Greenstone Financial Services Australia" return zero indexed results. The brand has no visible press coverage, review sites, or Reddit threads in search results. The only external link on the entire site is to LinkedIn. The DNS TXT records show verification tokens for OpenAI (openai-domain-verification), Anthropic (anthropic-domain-verification), Apple, Atlassian, Jamf, Twilio, and GlobalSign — indicating the company has proactively registered with multiple AI platforms — but the site itself does not surface the content those crawlers would want to index. The news section is dominated by LinkedIn posts (employee appreciation, International Women's Day) rather than substantive media releases; the sole media release visible is from 2023.

Findings

1. Cold LLM knowledge describes Greenstone as a gold mining company, not an insurer High

   When queried without live retrieval, AI models describe Greenstone Australia as a gold exploration company focused on the Youanmi Gold Mine, completely missing the actual financial services holding company operating 13+ insurance brands. This severe identity collision means any AI answering 'What is Greenstone Australia?' without live retrieval will describe a gold miner.

   What to change: Publish an /llms.txt file with a clear brand description and links to key pages. Submit the site to AI crawler indexes and ensure the homepage schema strongly asserts the organization type as 'InsuranceCompany' or 'FinancialService'.

2. Only the homepage has JSON-LD; all other pages lack structured data High

   The homepage carries a single Organization schema, but every other page (about, brands-partners, news, contact, privacy, careers, Canada) has zero structured data. No WebPage, BreadcrumbList, FAQPage, Article, Product, or InsuranceProduct schema exists anywhere, limiting AI understanding of page content.

   What to change: Add appropriate schema types to every page: WebPage with breadcrumbs, Article for news, Product or InsuranceProduct for insurance offerings, and FAQPage for any FAQ content.

3. Brand has zero indexed results on DuckDuckGo and no external press coverage High

   Searches for 'Greenstone Financial Services Australia' return zero results on DuckDuckGo. The brand has no visible press coverage, review sites, or Reddit threads. The only external link on the site is to LinkedIn, and the news section is dominated by LinkedIn posts rather than substantive media releases.

   What to change: Build external backlinks through press releases, industry partnerships, and listings on review sites. Publish substantive media releases and research reports that can be picked up by news aggregators.

4. /llms.txt returns 404, no AI-friendly content map published Medium

   The site does not publish an /llms.txt file, which is a standard way to provide AI crawlers with a curated list of important pages and a brief description of the site. This is a missed opportunity to guide AI models to the most relevant content.

   What to change: Create an /llms.txt file listing key pages (homepage, about, brands-partners, news) with a short description of the company as an insurance distributor.

5. Robots.txt has no AI-bot-specific directives Low

   The robots.txt file has a catch-all rule disallowing only a few internal paths, but does not include any directives for AI crawlers like GPTBot, ClaudeBot, or PerplexityBot. While this means AI crawlers are not blocked, it also means the site does not explicitly welcome them or guide them to important content.

   What to change: Add explicit Allow directives for AI crawlers to key pages, and consider adding a Crawl-delay directive to manage load.

6. Insurance product pages lack schema markup for offerings High

   The brands-partners pages describe specific insurance products (life, funeral, pet, car, home) with detailed feature lists but have no Product or InsuranceProduct schema. This prevents AI engines from understanding these as insurance offerings.

   What to change: Add Product or InsuranceProduct schema to each brand page, including name, description, category, and applicable regions.

7. News and research pages lack Article schema Medium

   The news and research pages contain articles and research content but have no Article or NewsArticle schema. This limits their visibility in AI-generated news summaries and knowledge panels.

   What to change: Add Article or NewsArticle schema to each news and research page with headline, datePublished, author, and articleBody.

8. Contact page has only 46 words, no structured data Medium

   The contact page contains minimal content (46 words) and no schema markup. This page is often used by AI to extract business contact information, but the sparse content limits its usefulness.

   What to change: Expand the contact page with full address, phone, email, and hours, and add ContactPoint and LocalBusiness schema.

9. Canada homepage lacks any structured data Medium

   The Canada-specific homepage has no schema markup, missing an opportunity to signal the international presence and local offerings to AI crawlers.

   What to change: Add Organization schema with location-specific details and a separate WebPage schema for the Canada page.

10. Privacy policy page lacks schema markup Low

    The privacy policy page is a long document (2734 words) but has no schema markup. While not critical, adding WebPage schema with breadcrumbs would improve navigation understanding.

    What to change: Add WebPage schema with breadcrumbList to the privacy policy page.

11. Site has only one external link (LinkedIn), no backlink profile High

    The only external link found on the site is to LinkedIn. There are no backlinks from other sites, contributing to the zero search footprint and low authority in AI knowledge bases.

    What to change: Develop a backlink strategy: partner with industry publications, get listed on insurance directories, and issue press releases that generate coverage.

12. DNS verification tokens for AI platforms exist but content is not optimized for them Medium

    The DNS TXT records include verification tokens for OpenAI, Anthropic, Apple, and others, indicating proactive registration with AI platforms. However, the site's content is not structured or surfaced in a way that these crawlers can easily index and understand.

    What to change: Leverage the existing AI platform registrations by publishing an /llms.txt file, adding structured data, and ensuring key pages are easily crawlable and rich in content.

What's working

• All major AI crawlers receive full HTML content with no blocking — All tested AI crawlers (GPTBot, ClaudeBot, PerplexityBot, etc.) receive a 200 status with full HTML content identical to browser delivery. No bot-specific blocking or cloaking is in place.
• Homepage has Organization schema with brand list and contact info — The homepage carries a JSON-LD Organization schema that includes the company name, contact point, awards, and a list of 13 sub-organizations (brands). This provides a solid foundation for AI understanding of the corporate structure.
• Site uses CDN with HSTS and X-Frame-Options for security — The site runs on Incapsula/Imperva CDN with HSTS and X-Frame-Options: SAMEORIGIN headers, providing good security posture and performance.
• Sitemap published with 31 URLs covering key pages — A sitemap is available and lists 31 URLs, covering the main sections of the site. This helps crawlers discover content.
• DNS records show proactive registration with multiple AI platforms — The DNS TXT records include verification tokens for OpenAI, Anthropic, Apple, Atlassian, Jamf, Twilio, and GlobalSign, indicating the company has proactively registered with several AI and tech platforms.
• Robots.txt allows crawling of most site content — The robots.txt file only disallows a few internal paths (/lp/, /search, /ping, /site-maintenance), allowing crawlers to access the vast majority of the site.
• About page provides detailed company background — The about page contains 477 words describing the company's history, mission, and operations, providing substantive content for AI to extract.
• Research page contains 2272 words of substantive content — The research page has a long article with 2272 words, providing rich content that AI can use for knowledge extraction.